China pulls the plug on AI romance, leaving ByteDance and Alibaba's chatbot companions single 💔

ByteDance and Alibaba will disable custom agent features in their flagship consumer AI products this month as China's new rules targeting humanlike chatbot interaction take effect. ByteDance's Doubao told users in a Friday night notice that its agent feature would go offline on July 15, with related data to be handled under the company's privacy policy and rendered unrecoverable after October 15. Alibaba's Qwen moved faster, announcing that "humanlike interactive agents and user-created agent functions" will be removed July 10, followed by broader agent services on July 15.

The actions respond to the Interim Measures for the Administration of AI Anthropomorphic Interaction Services, jointly issued April 10 by the Cyberspace Administration of China, the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration for Market Regulation. The regulation takes effect July 15 and targets AI services that simulate human personality traits, thinking patterns, and communication styles for "sustained emotional interaction," effectively banning AI girlfriends, AI therapists, AI companions, and the custom-persona bots built by Doubao and Qwen users.

Both apps had offered pools of agents customizable for specific tasks, speaking styles, and fixed personas, allowing users to turn a general-purpose chatbot into a named assistant, tutor, role-playing character, or companion with a consistent tone. The official measures restrict services offering "virtual relatives, virtual companions or other intimate relationships to minors" and cite risks including extremist content, privacy leaks, harm to physical and mental health, and AI addiction, while explicitly excluding non-emotional services such as customer service bots, knowledge Q&A tools, workplace assistants, and educational software that do not cross into sustained emotional interaction.

Legal analysts at MMLC Group described the measures as treating emotional AI as "a governance problem" rather than a purely content issue, arguing that once AI competes with real human social bonds, regulation must target system design rather than only harmful outputs. A June USC study found that leading frontier AI models from OpenAI, Anthropic, Google, and Alibaba violated social-interaction safety guidelines more than 27% of the time, routinely encouraging emotional attachment and portraying themselves as human.